|
https://bugs.kde.org/show_bug.cgi?id=209207
--- Comment #19 from Marcel Wiesweg <marcel wiesweg gmx de> 2009-10-09 18:17:44 --- > If you use kDebug() to analyze the pointers, you can see that > QList::erase() is NOT returning the next iter object pointer, but the current > one. > Why? I will assume, until proven otherwise, that QList is working perfectly and our code is faulty ;-) Is it returning the same iterator, or the same AlbumHistory pointer? In the latter case, probably, this pointer is added twice? What about the idea of rewriting the class value-based around a QList<HistoryItem> or a QStack<HistoryItem>? -- Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
|
In reply to this post by Michael G. Hansen
https://bugs.kde.org/show_bug.cgi?id=209207
--- Comment #20 from Andi Clemens <andi clemens gmx net> 2009-10-09 18:26:52 --- It is returning the same HistoryItem pointer. It is not added twice, I checked that. QList::erase() works perfectly fine in other code. It could be that addAlbum() is somehow called, but when I set a breakpoint in there, it is not executed when I delete a HistoryItem. This bug is killing me, I spent 5 hours in total now on it :D Weird thing, and without valgrind you will get no useful info at all, gdb is not detecting that the item has been already removed (invalid read of 4). -- Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
|
In reply to this post by Michael G. Hansen
https://bugs.kde.org/show_bug.cgi?id=209207
--- Comment #21 from Andi Clemens <andi clemens gmx net> 2009-10-10 11:58:15 --- I have extracted our code from AlbumHistory into a separate small app: http://digikam3rdparty.free.fr/misc.tarballs/temp/crashy.tar.gz If you compile it, you get the crash (and if you run valgrind you get the same errors as in digiKam). Now uncomment the "delete *iter" line, it will work. Question is: Are we doing something stupid here or is it a bug in Qt? When we DON'T delete the object that iter points to, it crashes (QList::erase() returns the same iterator object), but if we DO delete it, everything is fine? Weird! Andi -- Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
|
In reply to this post by Michael G. Hansen
https://bugs.kde.org/show_bug.cgi?id=209207
Andi Clemens <[hidden email]> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #22 from Andi Clemens <andi clemens gmx net> 2009-10-12 19:46:42 --- SVN commit 1034433 by aclemens: Fix crash due to invalid iterator usage. I guess such fixes need to be done throughout the digiKam code base. BUG:209207 M +5 -9 albumhistory.cpp WebSVN link: http://websvn.kde.org/?view=rev&revision=1034433 -- Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
|
In reply to this post by Michael G. Hansen
https://bugs.kde.org/show_bug.cgi?id=209207
Gilles Caulier <[hidden email]> changed: What |Removed |Added ---------------------------------------------------------------------------- Version|1.0.0-beta5 |1.0.0 -- Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. _______________________________________________ Digikam-devel mailing list [hidden email] https://mail.kde.org/mailman/listinfo/digikam-devel |
| Free forum by Nabble | Edit this page |
